Many diseases are now understood at the molecular and genetic level. Analysis of such molecules is important for disease diagnosis and prognosis. Previous methods for direct extraction of cellular tissue material from a tissue sample are limited because the extraction reflects only the average content of disease associated markers. In reality, tissues are very heterogeneous and the most diagnostic portions of the tissue may be confined to a few hundred cells or less in a lesion.
Normal tissue samples contain a variety of cell types surrounding and adjacent to the pre-invasive and invasive tumor cells. A region of the tumor tissue subject to biopsy and diagnosis as small as 1.0 mm can contain normal epithelium, pre-invasive stages of carcinoma, in-situ carcinoma, invasive carcinoma, and inflammatory areas. Consequently, routine scraping and cutting methods will gather all of these types of cells, and hence, loss of an allele will be masked by presence of a normal copy of the allele in the contaminating non-malignant cells. Existing methods for cutting away or masking a portion of tissue do not have the needed resolution. Hence the analysis of genetic results by those previous methods are always plagued by contaminating alleles from normal cells, undesired cells or vascular cells.
The molecular study of human tumors is currently limited by the techniques and model systems available for their characterization. Studies to quantitatively or qualitatively assess proteins or nucleic acid expression in human tumor cells are compromised by the diverse cell populations present in bulk tumor specimens. Histologic fields of invasive tumor typically show a number of cell types including tumor cells, stromal cells, endothelial cells, normal epithelial cells and inflammatory cells. Since the tumor cells are often a relatively small percentage of the total cell population it is difficult to interpret the significance of net protein or nucleic acid alterations in these specimens.
Studies of human tumor cells in culture do not account for the complex interactions of the tumor cells with host cells and extracellular matrix, and how they may regulate tumor cell protease productivity or activation. Immunohistochemical staining allows one to examine enzyme distribution in regions of tumor invasion, however, results vary with tissue fixation and antibody-antigen affinity, and provide only a semi-quantitative assessment of protein levels. Furthermore, quantitative interpretation of staining results is complicated by the variability of staining patterns within tissue sections, subjective evaluation of staining intensity, and the difficulty in interpreting the significance of stromal staining. In addition, many antibodies utilized in the study of proteases do not differentiate pro-enzyme from active enzyme species. Assays of enzyme or mRNA levels from homogenates of human tumors does not account for either the mixed population of cells within the specimens, or the concomitant pathophysiologic processes which may be occur in the tissue
Prior methods of study have not allowed investigators to specifically examine genetic alterations in pre-invasive lesions. Even the most sophisticated genetic testing techniques to date have been of limited value because the input DNA, RNA or proteins to be analyzed are not derived from pure cell populations exhibiting the disease morphology. Several methods have been reported for tissue microdissection to address this problem, including gross dissection of frozen tissue blocks to enrich for specific cell populations, irradiation of manually ink stained sections to destroy unwanted genetic material, touch preparations of frozen tissue specimens and microdissection with manual tools. These methods, however, are not sufficiently precise and efficient for routine research or high throughput clinical molecular diagnostic applications. Manual microdissection, for example, has good precision but is time consuming, labor intensive, requires a high degree of manual dexterity, and is not generally suitable for the ordinary technologist.
The present inventions provides a novel improved means to specifically examine genetic alterations in pre-invasive lesions of common epithelial tumors such as breast and prostate carcinoma. In particular, the present invention permits the microsampling of as few as one cell, with RNA and DNA extraction of the sampled cell. This method has been demonstrated to be extremely sensitive and to surpass previous and current technologies by more than two orders of magnitude. It has allowed the sensitive detection of loss of heterozygosity in early pre-invasive lesions being a gateway to the discovery of, for example, new genetic loci on chromosome 11 for breast cancer and a new genetic loci on chromosome 8 for prostate carcinoma.
The practice of the invention further permits the construction of genetic libraries from the extracted material. Thus, libraries from predetermined cells of interest, particularly abnormal cells, may be constructed and compared to libraries made from close-by, or adjacent, other cells, such as normal cells. Such libraries may be used, for example, to compare one or more specific genetic loci, the expression of one or more RNAs, particularly mRNAs, to isolate and/or clone one or more specific nucleic acid, and the like.